Bachelor's Thesis


Implementation of a hint system for solving Java programming exercises
Bachelor's Thesis, January 2015 — June 2015
Author: Baumann Cedric
Supervisors: Đurica Nikolić and Marco Piccioni


A challenge for Introduction to programming MOOCs (Massive Open Online Courses) is ensuring that students do not get stuck in solving the exercises when there are no teachers that can provide direct guidance. AutoTeach, our first attempt to tackle this challenge, is an incremental hint system for Eiffel programming exercises. Given a complete solution of a programming exercise, AutoTeach can generate several stripped-out versions of it, ordered by hint level, which initially only contain the skeleton of the solution's code, and then, for higher hint levels, gradually reveal more parts of the code and/or hints on how to write it. Students who get stuck on an exercise can ask for a hint (one or multiple times) and receive versions of the code where more and more parts of the solution are revealed, until they are hopefully able to proceed on their own.
The goal of this project is to develop jAutoTeach, an Eclipse plug-in which will implement and extend to Java the ideas of AutoTeach.